Research Article www.acsami.org
Quantifying Plasmon-Enhanced Light Absorption in Monolayer WS2 Films Serkan Butun,†,∥ Edgar Palacios,† Jeffrey D. Cain,‡,§ Zizhuo Liu,† Vinayak P. Dravid,‡,§ and Koray Aydin*,† †
Department of Electrical Engineering and Computer Science, ‡Department of Materials Science and Engineering, and §International Institute for Nanotechnology (IIN) Northwestern University, Evanston, Illinois 60208, United States
Downloaded via TU MUENCHEN on July 4, 2018 at 14:12:23 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
S Supporting Information *
ABSTRACT: Transition metal dichalcogenide semiconductors hold great promise in photonic and optoelectronic applications, such as flexible solar cells and ultrafast photodetectors, because of their direct band gap and few-atom thicknesses. However, it is crucial to understand and improve the absorption characteristics of these monolayer semiconducting materials. In this study, we conducted a systematic numerical and experimental investigation to demonstrate and quantify absorption enhancement in WS2 monolayer films, in the presence of silver plasmonic nanodisk arrays. Our analysis combining full-field electromagnetic simulations and optical absorption spectroscopy measurements indicates a fourfold enhancement in the absorption of an WS2 film near its band edge, close to the plasmonic resonance wavelength of Ag nanodisk arrays. The proposed Ag/WS2 heterostructure exhibited a 2.5fold enhancement in calculated short-circuit current. Such hybrid plasmonic or two-dimensional (2D) materials with enhanced absorption pave the way for the practical realization of 2D optoelectronic devices, including ultrafast photodetectors and solar cells. KEYWORDS: 2D materials, Transition-metal dichalcogenides, plasmonics, absorption enhancement, WS2
1. INTRODUCTION In the bulk, transition metal dichalcogenides (TMDCs) are composed of many semiconducting monolayers stacked on top of each other that are weakly bound by van der Waals forces. Such weak bonding between adjacent layers has enabled the isolation of monolayer films using a variety of exfoliation methods.1−5 On the other hand, more recent advances in synthesis using chemical vapor deposition (CVD)6−9 and molecular beam epitaxy10,11 have allowed researchers to further explore electronic transport and the optical properties that arise from the delocalization of electrons and band-structure realignment.5,12 By taking advantage of these desirable properties, two-dimensional TMDCs (2D-TMDCs) have been proposed and demonstrated for many high-performance applications, including high-speed transistors,13−19 photodetectors,20−23 emitters,24,25 and valleytronic devices.26 Although monolayer 2D-TDMCs possess advantageous properties that are potentially useful in the state-of-the-art optoelectronic applications, such as direct band gaps and high mobility, they also pose a major challenge in light−matter interactions due to their inherent mono- to few-layer thicknesses. Light absorption and emission strongly depends on the material volume. Although monolayer TDMCs have excellent optical properties, such as high absorption coefficient and PL emission, they exhibit less than ideal absorption © 2017 American Chemical Society
characteristics due to their thickness, hindering their ability to be deployed in practical optoelectronic applications. Strategies that enhance light−matter interactions and boost absorption and emission in 2D-TMDCs would provide huge impact toward the practical realization of true 2D optoelectronic devices. To address this challenge, many groups have attempted to utilize different resonator configurations to confine light within the 2D-TMDCs to increase emission and absorption.27 Thus far, several designs have been utilized, such as distributed Bragg reflector microcavities28,29 and plasmonic nanostructures with periodic30 and nonperiodic elements.31 Previous studies were concentrated mostly on enhancing light emission, which is an indication of enhanced light absorption. In addition, some studies have shown promise for enhancing absorption by demonstrating enhanced photoconductivity31 through the dispersal of plasmonic nanoparticles on rather small-area MoS2 flakes. A small number of studies have also shown the potential of TMDCs as an optoelectronic material by studying the absorption of the few-layer structures.21,32,33 However, a thorough study of the absorption properties of single-layer 2D materials has not yet been reported, as it often entails large-area Received: February 9, 2017 Accepted: April 10, 2017 Published: April 10, 2017 15044
DOI: 10.1021/acsami.7b01947 ACS Appl. Mater. Interfaces 2017, 9, 15044−15051
Research Article
ACS Applied Materials & Interfaces
Figure 1. (a) Conceptual schematic of the proposed WS2/Ag plasmonic heterostructure on sapphire. (b) Sample Raman spectrum of the monolayer WS2 film on sapphire substrate. (c) Two-dimensional atomic force microscopy (AFM) image of one of the individual WS2 flakes. A linecut over the edge is superimposed on the picture to indicate the relative height difference. The scale bar is 5 μm. (d) Dark-field optical micrograph of the fabricated plasmonic Ag arrays with periodicity of 300 nm. Diameters of Ag nanodisks in each square: bottom left to right 75, 90, and 100 nm; top left to right 115, 130, and 155 nm. Each square is 50 μm wide. (e) Corresponding 2D photoluminescence (PL) measurement of the same area shown by the yellow outline in (d). (f) Scanning electron microscopy image of one of the plasmonic arrays on WS2. Areas where there is no WS2 appear darker; the yellow triangle indicated one such area. The scale bar is 1 μm.
inherently require a transparent substrate. Figure 1a shows a schematic drawing of the Ag/WS2 heterostructure. We have fabricated Ag nanodisks directly on the WS2 film using electron-beam lithography. The direct band gap of monolayer WS2 is 620 nm.34 Thus, Ag is a suitable metal for exciting plasmonic resonances at both shorter and longer wavelengths than the band gap wavelength of WS2 by varying the diameter and periodicity of the plasmonic nanoparticle arrays.35,36 We performed Raman scattering measurements on our WS2 film to confirm the number of layers and the material quality. The positions of the peaks observed at 297 cm−1 (2LA-2E22g), 324 cm−1, 352 cm−1 (2LA-E12g), and 417 cm−1 (A1g) are consistent with the values reported in the literature for monolayer WS2 films.37,38 There is an additional sapphire peak at 417 cm−1, which is convoluted to the A1g peak of WS2. To verify the number of layers of WS2 films, we carried out AFM measurements. Figure 1c confirms the step height of 1 nm at the edge of our large-area WS2 film, which is an indication of monolayer thickness.39 The area of the monolayer WS2 film extended over millimeter size on sapphire substrate, which is confirmed by visual inspection in both optical
materials. In addition to large-area materials, a transparent substrate is required to measure the absorption spectra. Almost all reported studies were performed on Si/SiO2 substrate, which makes spectral transmission measurements in the visible region impossible. In this study, we demonstrate how periodic Ag plasmonic nanodisk arrays are able to confine light within monolayer-thick 2D-TMDCs for enhancing absorption. By conducting both finite-difference time domain (FDTD) electromagnetic simulations and visible-frequency optical spectroscopy, we deconvolute and quantify the absorption inside the WS2 film and demonstrate that the Ag/WS2 heterostructure can provide a fourfold absorption enhancement. The results presented here do not only hold promise for absorption enhancement in WS2 monolayers but could also be used along with other 2DTMDCs, therefore making it a versatile design.
2. RESULTS AND DISCUSSION Our proposed structure consists of Ag nanodisks fabricated on a CVD-grown monolayer WS2 on a double-side-polished sapphire substrate because optical absorption measurements 15045
DOI: 10.1021/acsami.7b01947 ACS Appl. Mater. Interfaces 2017, 9, 15044−15051
Research Article
ACS Applied Materials & Interfaces
Figure 2. Measured spectral (a) reflection and (c) transmission data of the fabricated Ag/WS2 heterostructures as a function of diameter. (b, d) Corresponding FDTD simulations of the proposed structures.
attributed mostly to the grainy nature of the polycrystalline WS2 film and missing small patches. These variations are greatly enhanced in Ag-array regions because of the enhanced emission contrast. Six square high-PL-intensity regions in Figure 1e correspond to the enhanced PL due to the interaction between localized surface plasmon resonance of Ag nanodisks and WS2. Similar structures on TMDCs30,40 have also shown similar PL enhancements, in which the PL enhancement mechanism is thoroughly discussed. A high-magnification SEM image of one of the fabricated arrays is shown in Figure 1f. It is important to note that these structures are fabricated on an insulating substrate; therefore, the image quality is not as good as that of similar structures that are usually fabricated on Si substrates. Methods section provides a detailed description of how we acquired SEM images on insulating substrate. In addition, it is well known that Ag oxidizes very rapidly. This would have a significant effect on final results. In our experiments, significant shifts were observed after 2 weeks of fabrication. Therefore, we performed all of the measurements within 2 weeks after fabrication. We also purposely delayed SEM images to avoid ebeam damage. SEM images were acquired about 2 weeks after the fabrication. Hence, a slight misshaping and/or warping of Ag nanodisks manifested itself due to oxidization. Nevertheless, these images contain ample information, such as some small patches of missing WS2 film (indicated by the yellow triangle). We attribute the variations of the PL intensity across the surface to these voids, as mentioned above. We have performed spectral reflection and transmission measurements of fabricated Ag/WS2 heterostructures. In Figure 2, we plot reflection (a) and transmission (c) spectra of hybrid Ag nanodisk arrays/WS2 material for three different nanodisk diameters: 90, 100, and 115 nm. A dip at 620 nm, which is the band gap wavelength of WS2, is evident in all reflection and transmission spectra. We explain this dip by the superimposed absorption of WS2 over the localized plasmon resonance of the Ag nanodisk array. In addition, a redshift in the resonance peak wavelengths with increased diameter is clear in both the
microscope and by contrast SEM images. Apart from very small voids (∼1 μm), which can only be seen in SEM images, the polycrystalline WS2 layer was uniform and continuous. The contrast difference between the WS2 film and the sapphire surface can be seen in the SEM image highlighted by the yellow dashed triangle in Figure 1f. These voids are left after the CVD growth of the WS2 film. In addition, film uniformity was confirmed by 2D PL scans, which we discuss in detail later in this article. Such large area of monolayer 2D material enabled us to perform an iterative investigation of plasmonic nanodisk arrays as a function of structural parameters, such as periodicity and diameter. Considering that, here, we investigated the effect of plasmonic resonance tuning on the optical absorption of WS2 films by changing the diameter of Ag nanodisks, using individual flakes would make this work unfeasible as each plasmonic array would require a sufficiently large and pristine flake. Therefore, Ag nanodisk arrays with various diameters all having a periodicity of 300 nm were fabricated on the large-area monolayer WS2 film. In addition, nanodisk arrays were placed in close proximity to each other, to eliminate spatial variance in film quality. The micrograph in Figure 1d shows one set of plasmonic arrays of Ag nanodisks with diameter varying from 75 to 155 nm. Each array is 50 μm wide. The different color of each square is a manifestation of the localized surface plasmon resonance. We should note that because of severe adhesion problem, some of the Ag nanodisks are missing (
